Preservation of Digital Content Jacob Nadal Head, E. Lingle Craig Preservation Laboratory Indiana...

Post on 16-Dec-2015

214 views 1 download

Transcript of Preservation of Digital Content Jacob Nadal Head, E. Lingle Craig Preservation Laboratory Indiana...

Preservation of Digital Content

Jacob NadalHead, E. Lingle Craig Preservation LaboratoryIndiana University Bloomington

jnadal@indiana.edu(812) 855-6281

•“The first principle of digital preservation is: Decide at the time when it is created how long the material is to last” -- Susan Keene

•Most library preservation, however, is the rescue of items that have not benefitted from this sort of forethought.

•Digital presrvation, as Keene describes it, more closely aligns with conservation through control of environmental factors, use conditions, and pre-treatments, such as mass deacidification.

Overview

•Traditional Branches of Preservation

•Example: Linear B

•Managing and sustaining bits

•Making bits useful

•Example: The BBC Domesday Book

Traditional Branches of Preservation

•Conservation

•General Collections

•Special Collections

•Reformatting

Conservation•Conservation serves to increase the

usability of the original item.

•Special collections attempts to maintain the state of original issue.

•General collections attempts to improve upon the state of original issue.

•Conservation seeks to create an environment and use conditions that promote to longevity, or minimize decay.

Special Collections Conservation

•Individual artifacts are maintained as unique examples of their kind.

•When changes are made, they are either:

•Made in close accord with the original terms of production, or they are

•Unobtrusive alterations to make the item usable under carefully defined conditions.

General Collections

Conservation•Individual items are maintained in the same format, but with changes made to:

•Increase longevity and durability of the item, or

•Increase the speed of treatment.

•A book remains a book, but the components of that book may be replaced or augmented.

A damaged book with

intact hinges

The original case was discarded and the text block sewn into a new case, using cloth hinged double end sheets. This volume was then hit by a motor vehicle. Despite severe damage, the hinges made by the general collections conservation unit are intact. The original format, with improved durability.

Damaged book, Intact hinges

Damaged book, Intact hinges

Reformatting

•Retains information by copying “data” from failed media to a more durable media.

•Entails loss of artifactual value.

•Unlike decision between types of conservation, reformatting is required by condition, rather than intention of use.

Microfilm

•Our first stab at the digital library, or at least the library of the future:

•Compact,

•Durable,

•Easily reproduced, and

•Easily distributed, but...

Never very popular.

•Changed method of interaction - from “leafing” to “scrolling.”

•Most useful for bi-tonal text and line art.

•And no one ever did build that Memex...

•And increasingly, no one is building microform equipment.

Audio-visual crisis

•Media is inherently fragile

•Reformatting requires original playback equipment

•Often obsolete or in poor repair

•Difficult to integrate with newer recording and playback technologies.

Audio-visual crisis

•Because information is recorded as machine readable data, maintaining the media is not enough.

•Playback equipment is required.

General Principles in Preservation

•All three preservation approaches are concerned with predicting user needs.

•If intellectual value is in the object, then conservation is the only recourse.

•If intellectual value is encoded in text, then reformatting or collections conservation are viable.

General Principles

•Perpetuating the object has served well as long as the information encoding is human readable.

•With machine readable data (optical discs and magnetic media), preservation of the use environment has become necessary.

Linear B

•Depended on existence of encoded data on a durable medium - analogous to traditional preservation.

•Data was useless without de-coding - analogous to a/v and digital presrvation.

Linear B•Bronze Age Cretan script: c. 1450 to

1375 B.C.

•No cribs, such as the Rosetta Stone, an almost entirely logical decipherment

•Discovered by Sir Arthur Evans, in spring of 1900 on numerous inscribed clay tablets.

•Largely inventory data

First successes•Counting system was easy to

determine

•Writing direction from left to right

•90 distinct characters, indicative of a syllabic system

•Debate over relation to Greek or Cypriot. Most felt it was a unique Cretan language.

Alice Kober - the hard work begins.•1940 - Alice Kober identifies word

triplets

•Same word stem with different endings, presumably for case (e.g. accusative, or nominative)

•Kober separated symbols into modifiers and word stems

Michael Ventris - from patterns to

prose•Consonant-vowel patterns established

•Problem of missing vowels and leading vowels: e.g. di-vi-si-b(i)-le or i-n(i)-di-vi-si-b(i)-le

•Developed refinements of Kober’s chart to manage these relationships

A few good guesses

•Refinement of relationships gave Ventris enough confidence to take a guess at three words, the towns of Anisos, Knossos, and Tulissos

•Assigning consonant values opened up more words

•Greek philologist John Chadwick partnered to carry forward the decoding of a Greek dialect from the time of the Trojan War.

Linear B•Began with identification of

recognizable parts - could be blocks on a disc or triplets of words.

•Located correspondence to possible informational content (names of towns, etc...)

•Required some subject expertise to instantiate words.

From

Book and Paper to

Digital

Resources

Why use digital means for

preservation at all?•Digital content is immune to daily

wear and tear, mold, or insects

•Ability to duplicate across many sites provides means for protection from local disaster

•Many important materials are now born digital.

Methods of Digital Preservation

•Bit preservation: Transfer and refresh of data

•Migration: Transformation of data into new formats to allow for continued access

•Emulation: Recreation of original operating environment for continued access

•Conservation: Maintaining original equipment for access

Some examples of things to keep...

Aligning traditional to digital

preservationSpecial

Collections

Unique or rare items, and items with

importance as artifacts

Museums, Rare books libraries,

Archives,

Joshua Davis - http://www.joshuadavis.co

m

King Kipple - http://www.funnygarbage.

com/kingkipple/

General Collections

Common issue items, with primarily

intellectual value

Research collections

WilsonWeb - http://vnweb.hwwilsonweb.com/hww/shared/shared_

main.jhtml

Project Gutenberg - http://gutenberg.net/

ReformattingItems that are

damaged beyond use

Anywhere and everywhere -

not an intellectual

class

BBC Domesday Book

GPD Floppy Disk Project - http://www.indiana.edu/~libgpd/mforms/floppy/floppy

.html

Aligning traditional to digital

preservationSpecial

Collections

Maintaining the original or copy of record

Maintain “look and feel” of original use.

Maintain original

equipment.

Emulation and Preservation

General Collections

Maintaining original item by improving usability and

durability

Maintain the content of the

resourceMigration

ReformattingCreating

facsimile of the original

Maintain content or, at a minimum, bit-stream of the original

Migration

Preserving Bits

•Storage Media

•Magnetic Tape

•Magnetic Disc

•Optical Disc

•Media independent storage

Magnetic Tape

•High data density, and low cost

•Quantum gives a life-span of 30 years for its DLT IIIxt and IV at 68º F (20º C) and 40% non–condensing humidity.

•IBM gives a life-span of 30 years for its LTO Ultrium tapes at 61-90° F (16-32° C); 20-80% RH, with a 79° F (26° C) wet bulb max

Magnetic Disk•Common MTBF ranges are from

500,000 to 1,200,000 hours (57 to 137 years), meaning half the drives in the population will fail before that time, half after.

•Seagate and Maxtor offer 5 year warranties on their SCSI HDs

•High probability of mechanical damage.

Expected Survival Rates for Hard

Drives500,000 750,000 1,000,000 1,250,000

3 94.88% 96.56% 97.41% 97.92%

5 91.61% 94.33% 95.71% 96.56%

10 83.93% 88.98% 91.61% 93.23%

15 76.89% 83.93% 87.69% 90.02%

20 70.44% 79.17% 83.93% 86.92%

25 64.53% 74.68% 80.33% 83.93%

30 59.12% 70.44% 76.89% 81.04%

Years of Use vs. Mean Time Between Failures (MTBF)

Anticipated Life-span of Hard Drives

% su

rviv

i ng

Years of Operation

MTBF Rating

Optical Disc

•CDs and DVDs are multilayered products, consisting of polycarbonate outer coatings (plastic), metal reflective layers, and dye layers

•Life-spans range from 3 - (predicted) 300 years, depending on components

Dye Layer (1st Data Layer in Recordable DVDs)

Optical Disc LayersPolycarbonate Layer

Metal Reflective Layer (Data Layer in Manufactured CDs)Dye Layer (Data Layer in Recordable CDs)

Polycarbonate Layer

Polycarbonate Layer

Metal Reflective Layer (1st Data Layer in Manufactured DVDs)Dye Layer (1st Data Layer in Recordable DVDs)

Polycarbonate Layer

CD-R or CD-R/W

Dual Sided DVD±R or ±R/W

CD Metal Reflective Layers

Metals Properties and Problems

AluminumLeast reflective, subject to

tarnishing.

SilverMost reflective, relatively safe

from tarnishing. LE 100.

GoldSlightly (5%) less reflective than

silver, but very unlikely to decay. LE 300.

Common CD DyesDye Color

Properties and Problems

Azo Very deep blue

Very photosensitive. Failure after 20 –100

hours of full spectrum light

exposure.

Cyanine Blue

Photosensitive. Failure after 100 – 200 hours of full

spectrum exposure.

Pthalocyanine Very light greenFairly stable. Failure after 500 – 600 hours

of exposure.

Optical Disk for Archival Storage

•Some formulations may be viable for long term storage

•Require numerous disks to provide sufficient capacity (1,614 CDs / TB)

•Time consuming to read back data for large collections.

Media independent

storage•Takes advantage of a fundamental

property of digital content, to create identical copies on changing media.

•Media is cycled out before end of life.

Digital archive systems

•Manage Refresh and Transfer of data

•Refresh: Copying the data to a newer carrier of the same type (Tape to Tape)

•Transfer: Copying the data to a more stable carrier (Hard Drive to Tape)

•Require significant, ongoing, commitment of staff and resources.

The IU MDSS

•Hierarchical Storage Management System (HSM)

•Disk caches of roughly two terabytes, back-ending into two tape libraries which provide a total uncompressed data storage capacity of nearly 1.6 petabytes.

Transfer and Refresh in the

MDSS•Step 1: Data written to disk cache

•Step 2: If not accessed, data is transferred to tape

•Step 3, 4, 5... Data on tape is refreshed onto new tapes.

OCLC Digital Archive

•Currently provides bit preservation service

•Ingest involves creation of a METS compliant SIP, in accordance with the OAIS model.

•No current facility for migration or emulation.

General Principles•Most media used for digital storage are

not robust enough for archival purposes.

•Requirement for perpetuation of data depends on media independent storage

•Media independent storage requires an ongoing commitment of staff and facilities to manage transfer and refresh functions

Making

Bits

Useful

Applications and Operating Systems

•Operating system manages storage and retrieval of data, parts of the display funtion, and distribution of different data types to appropriate applications

•Applications interpret data for display and manipulate data.

Preservation approaches

•Migration manipulates data to make it usable by different applications and operating systems.

•Emulation recreates applications and operating systems to run original data.

Migration

•Porting or modifying the data into a more recent or widely accepted format

•Open Archiving Information System (OAIS): refreshment, replication, repackaging and transformation

Migration Variations

•Minimum migration: example could be a word processor file that is stripped of all but the ASCII characters. This simpler format is then migrated forward or refreshed as needed.

Migration Variations

•Migration on request: Proposed by CAMILEON Project

•Maintains original bit-stream and renders into a new format on demand, rather than through sequential transformations

Examples of migration projects•The Preserving and Accessing

Networked Documentary Resources of Australia (PANDORA)

•Migration of HTML, moving deprecated tags to match current specification

•CAMILEON: Vector Graphics Migration on Demand Tool

Emulation

•Uses software to emulate different software, hardware, or operating systems

Examples of emulation projects

•Networked European Depository Library (NEDLIB) - packaging digital resources with information about original use environment for later emulation

Emulation projects

•CAMILEON: BBC Micro Emulator as part of Domesday Project

And that’s before you just try Google...•List from: http://www.emulator-zone.com

•465 other emulation sites at: http://dmoz.org/Computers/Emulators/

More emulators

And then there’s plain old fashioned

conservation.•Preserves computer systems as

objects

•Complicated by object’s inherent fragility

Computer History Museum

•http://www.computerhistory.org/

•1401 N. Shoreline Blvd., Mountain View CA 94043

•25 years of collecting

•Collection now contains more than 4,000 artifacts, 10,000 images, 4,000 linear feet of cataloged documentation, and gigabytes of historic software.

Open Source and Standard Formats•Building systems with open source

software and open standards for files is emerging as a crucial preservation activity

•Allows for known starting points and migrations paths, and ensures basic information needed for emulation is available.

BBC Domesday Book

•Digitally captured census of the British Isles, after the Domesday Book of 1086.

•Ran on BBC Micro, Sinclair Spectrum, Research Machines

•Stored data on Laser Disc, what came to be known as LVROM

•As of today only 2 or 3 known systems left in operation.

BBC Domesday Book•Two discs: Community Disc and National

Disc

•National Disc contained a VR gallery as an index to the content

•Navigation used a trackball or touchscreen

•Community Disc was nvagated by a map interface

•National Disc was navigated by the VR gallery, map interfaces, or text searches

Migration and Emulation

•Two efforts:

•Migrate data to new format

•Emulate original operating environment

Emulation

•CAMELION

•Emulation of original operating environment under Windows

•Includes most of the 3 interactive sides, including overlay map data for National Disc and 3d Walkthroughs

Migration

•UK Public Records Office

•Uses broadcast video technology to store data

•Copying master 1” C format video to D3

•Then from D3 to Digital BetaCam using Transform decoder.

Domesday Successes

•Emulation: Show to be possible, but still incomplete

•Migration: Show to be successful, but already a “medium-term” solution, with eventual transfer to another digital format expected in 20 years.

And the big

Domesday Problem...

RIGHTS

Rights (sorry)

•Many of our most heavily used resources are under strict rights protection

•Even if content is not protected, interface to that content is (e.g. Gov’t info on Lexis-Nexis)

Our traditional rights

•Libraries have right to make copies of material for preservation purposes under section 107, traditionally meaning:

•1 copy for safe deposit (access restricted to library)

•1 copy for duplication (access restricted to library or designee)

•1 copy for use (user accessible)

Solutions?•Dark archives: digitally preserve

licensed content in closed spaces, until such time as rights change or copyright expires

•LOCKSS has such a facility in its architecture

•New licensing agreements

•Encouraging e-resource publishers to follow digital library standards and practices

What’s in place at IU•Bit preservation: MDSS

•Content management: OAIS, DLXS, FEDORA

•LOCKSS Participation

•Migration-capable: XML encoded documents, Standards-based projects.

•Many DLP projects have used migration techniques to create derivatives of a master file for on-line display

Major Trends

•Creation of bit storage systems

•Development of standards for encoding

•Promotion of open source solutions

•Development of roles for migration and emulation

Issues

•Nature of a “digital object” - how to best package data, runtime environments, and associated metadata

•Appropriate applications of migration and emulation strategies

•Rights issues related to the preservation licensed content

Projects• CEDARS: http://www.ukoln.ac.uk/metadata/cedars/

• CAMILEON: http://www.si.umich.edu/CAMILEON/

• interPARES: http://www.interpares.org/

• NARA ERA: http://www.archives.gov/electronic_records_archives/

• NDIIPP: http://www.digitalpreservation.gov/

• NEDLIB: http://www.kb.nl/coop/nedlib/

• OCLC: www.oclc.org/digitalarchive/

• PANDORA: http://www.nla.gov.au/policy/plan/pandora.html

• PRISM: http://www.prism.cornell.edu/

Works Consulted:• Beagrie, N (2003). National Digital Preservation Inititatives: An Overview of Developments in Australia, France, the

Netherlands, and the United Kingdom and of Related International Activity. Washington, DC: Council on Library and Information Resources.

• Chadwick, J. (1990).The decipherment of linear B. Cambridge: Cambridge University Press.

• Chapman, S. (2003). "Counting the Costs of Digital Preservation: Is Repository Storage Affordable?" Journal of Digital Information 4(2).

• Council on Library and Information Resources. (2002) Building a National Strategy for Digital Preservation: Issues in Digital Media Archiving Washington, DC: Council on Library and Information Resources..

• Digital Library Federation (2002). "Benchmark for Faithful Digital Reproductions of Monographs and Serials."

• Hunter, G. S. (2000). Preserving digital information: a how-to-do-it manual. New York, Neal-Schuman Publishers.

• Lavoie, B. F. (2003). "The Incentives to Preserve Digital Materials: Roles, Scenarios, and Economic Decision-Making."

• Lawrence, H. A. (2003). "The Need for a Reference Archive." E-Doc May/June 2003: 36-37.

• Rogers, M. (2003). "OCLC Delving Deeper into Digital Preservation Services." Library Journal 128(9): 27-28.

• Rothenberg, Jeff. (1999). Avoiding Technological Quicksand: Finding a Viable Technological Foundation for Digital Preservation. Washington, DC: Council on Library and Information Resources.

• Singh, Simon. (1999). The Code Book: The Evolution of Secrecy from Mary Queen of Scots to Quantum Cryptography. New York : Doubleday.

• Washington State Library. Washington State Library Digital Best Practices. Available at (Accessed Mar 13, 2004): http://digitalwa.statelib.wa.gov/newsite/best.htm